Skip to main content

Why we’re talking more about how we collect and analyse data on ethnicity

Today the Race Disparity Unit is publishing the first in a new series of methods and quality reports. The first report is of an overarching nature - it is our Quality Improvement Plan, which outlines the limitations of ethnicity data that we’re aware of, and how we plan to address them.

Computer screen showing the front page of the Race Disparity Unit Quality Improvement Plan

The benefits for our users (and ourselves)

We hope publishing these reports will benefit both our users and our own work in 3 ways:

First, the reports will help our users to understand issues that might affect how they interpret our data. For example, the report about ‘relative likelihoods’ will emphasise the importance of considering whether they are statistically significant, as well as how large they are, and explains how to go about calculating significance.

Second, they will help us improve our website if they highlight ways we can better collect, analyse, and publish ethnicity data. For example, our report on the ‘Mixed and Other’ ethnic groups will highlight the variation within these aggregated groups in the information we present for them on Ethnicity facts and figures.

Third, they will help us reach out to statisticians and other analysts who might be facing similar issues (related to either ethnicity or technical statistical issues). We can contribute to developing both the community and good statistical practice.

What we have planned

We are planning to publish reports around 3 themes in the coming year.

General improvements to ethnicity data

These reports will discuss:

  • the challenges of harmonising ethnicity classifications across government and other sources of data
  • the statistical implications of clustering different ethnic minority groups in particular areas, including sample designs, for example
  • our work with the Office for National Statistics (ONS) on linked data - setting out how we might use linkage, (i) to understand or to tackle quality limitations, including comparing ethnicity reporting between different data sets, and (ii) to generate new data to fill gaps or to provide fresh insights that aren’t available from a standalone data set

Data about particular ethnic groups

These reports will discuss:

  • data that is available for the ‘Mixed and Other’ ethnic groups, whether the groups are more similar or diverse in their characteristics, and whether a consistent set of reports should be produced for each of the groups
  • what we do about people whose ethnicity isn’t known, and the challenges this brings

Informing the use and interpretation of ethnicity data

These reports will discuss:

  • how we use relative likelihoods to understand how big disparities are between different ethnic groups
  • the impact of geographical distributions on the presentation and interpretation of police stop and search data, and the use of 2011 Census population estimates as denominators in the calculation of stop and search rates
  • how third party ethnicity data compares with self-reported data, providing examples from the criminal justice system
  • how many years of data it’s meaningful to combine, which will help to inform our practice of combining multiple years’ data to enable us to analyse smaller ethnic groups

Get involved

Email us or leave a comment below if you are interested in working with us on any of these topics, or have ideas about other methods and quality report topics.

Sharing and comments

Share this page